It is important for a business to maintain a website in some form or another. Some business websites are purely informational, providing basic details on the company. Others are a means to gain new leads or sell products. Still others could even be considered the business’ product itself, such as media outlets specializing in the creation and distribution of digital content. No matter the primary purpose, all websites benefit from a cycle of continuous, data-backed improvements, and fortunately, there is no shortage of data to be analyzed in the digital realm. A key to successful analysis is choosing the appropriate tools to gather, process, and analyze the data. Google provides two widely-used tools to help accomplish these tasks—Google Analytics and BigQuery.

Google Analytics Provides Insights into Website Usage

One of the most popular and accessible tools for website analysis is Google Analytics. Google Analytics is a tool that tracks and collects website data and provides an interface to build reports and run analyses on users’ website activity. This tool can track a wide variety of data surrounding user interactions, such as whether a user is new or returning, frequency of usage, session duration and bounce rate, traffic or ad source, device and browser used to access the site, geographic location, pages visited within a session, event tracking, ecommerce data, and many more, which are available through reports via the Google Analytics web interface.

However, much of the data available through the Google Analytics interface is presented in aggregate. Aggregated information can be helpful when analyzing overall trends, but limiting when trying to gain a deeper understanding of user behavior on an individual level, such as a clickstream analysis (clickstream refers to the chronological record of pages visited and click events, if configured, in a session). One solution is to export the available data records into a database system though the Google Analytics API, allowing data scientists to run more queries and use advanced modeling techniques to get deeper insights.

Unfortunately, Google Analytics limits the amount of data that can be exported through its API. The API only allows data scientists to export a limited number of metrics and a limited record of interactions. Websites with large amounts of traffic are only able to export sampled sets of the data rather than the full datasets, which leads to two main problems. First, the smaller amount of metrics available limits the analyses able to be performed and increases difficulty. With big data, the more data, the better when it comes to advanced analytics. The second issue is that the sampled dataset makes it difficult to target users on an individual level. The activity provided in the sample may not be representative of an individual’s browsing habits since it may contain only some of their sessions, or none at all. In addition, fewer users are linked to individuals in other data sources, such as a company’s internal customer or subscriber databases, because they may or may not be included in the sample. This leads to “holes” in the data record.

Using BigQuery to Go Deeper

Google does offer a solution to this dilemma: an export from Google Analytics Premium to BigQuery, Google’s cloud database solution. This method allows data scientists to export the full, unsampled record of data collected by Google Analytics, and take advantage of the increased processing ability of the BigQuery database service or their own internal database systems. The full dataset gives a more accurate view of individuals’ browsing habits, with fewer “holes” in the data. Additionally, more individuals from internal data sources—such as website or email subscribers, existing customers, or sales prospects—can be linked to website browsing records. Businesses can use this data to customize their website to more closely meet user (and customer) needs, with the ability to track changes and run experiments for continual quality control and improvement.

Not only is the data unsampled and served up in its entirety, but with an export to BigQuery, data scientists have access to over 100 available fields per click; for every page, event, transaction, or app interaction that has been configured. This allows them to view each interaction from many more angles than through Google Analytics alone, making more advanced analytics methods possible.

Conclusion

Google Analytics is a useful tool for answering a wide range of web analytics questions. However, as a business progresses along its analytics journey, it may be necessary to enlist additional tools in order to use more advanced queries and gain a deeper view of the data. This can be achieved by exporting data from Google Analytics Premium to BigQuery. The result is that the available data for analysis grows in both breadth and depth; the breadth of data increases since every interaction is available rather than only a sampling, and the increased number of dimensions available for analysis means a greater depth of analysis. Data scientists can then create customized algorithms to answer more detailed business questions, ultimately leading to insights that can improve the performance of the website in order to reach business goals faster and more efficiently.